A recent Agency for Healthcare Research and Quality (AHRQ) executive summary indicated that better systems are needed to determine the relative merits of existing versus new and expensive biologic drug therapies for rheumatoid arthritis (RA). There is currently no clear paradigm for how these different biologic therapies should be used in the clinic. CCE studies and biomarker predictions are needed to provide rationale algorithms for biological therapy selection given the extensive array of therapies for treating RA. There are now 8 (a ninth in the pipeline) expensive biological therapies approved by the FDA to treat RA. Research that addresses the comparative and cost effectiveness (CCE) of drug therapies is hindered by several factors. On the one hand, CCE studies that rely on randomized drug trial data do not provide the best estimates of real world health care costs and there is significant disparity between CCE results from randomized versus observational studies in RA patients treated with biologic therapies. However, the absence of randomization in observational studies imposes limitations on the analysis of CCE data. For example, the empiric and non-random selection of therapies for patients in the real world is complicated by confounding or selection bias that significantly impact the outcome of any CCE analysis including differences between groups in disease severity, medication compliance, and subject distribution into different treatment categories. Finally, CCE research in RA is hindered by a lack of biomarkers that predict responsiveness to one therapy versus another. The ideal system for CCE research would allow real world costs to be captured in patients that were randomly assigned to comparable therapies. We propose a Novel approach to address these current obstacles for performing not only CCE research but also streamlining efforts for personalized medicine. Essential to this concept is that following randomization, all other aspects of care would be governed by the patient and their physician, including decisions about drug dosing, monitoring and discontinuation. Randomization is needed to distribute patients evenly and remove biases, and real world observation is needed to capture accurate information on actual costs and therapeutic effectiveness. We refer to this study design as randomized observation. We plan to overcome the barriers to CCE research in RA by utilizing the University of Pittsburgh Medical Center (UPMC) RA Comparative Effectiveness Research (RACER) system to perform randomized observational studies to compare real world effectiveness of different treatment strategies. Initially, we will obtain CCE data from collaborators at Harvard and also from the UPMC RACER. The UPMC RACER system will utilize a large network of UPMC rheumatologists that are already linked by an electronic medical record (EMR) system;the EMR will be used to identify RA patients and to capture information on treatment, medical costs and clinical laboratory data. In conjunction with the analysis of over 1,110 RA patients followed at Harvard in the BRASS registry, we will demonstrate the UPMC RACER system's utility in an analysis of biologic therapies for treating RA and we will use the results of this analysis to design a future randomized observation study of biologic therapy for RA. This project involves a collaboration with researchers at Harvard University and the University of Pittsburgh with the goal of establishing the systems in order to effectively perform real-world cost-effectiveness research in patients with RA. We will also collect biological specimens for analyses of mechanistic studies and potential biomarkers so we can also provide data that potentially will be applicable for additional studies that will focus on research that will lead to personalized medicine for patients with RA.